Why do only some female swallowtails in Japan wear a disguise?
Catch two female swallowtail butterflies of the exact same species in the same Japanese field, and you might be looking at two completely different animals: one with the ordinary black-and-yellow tiger-stripe pattern of its species, the other with bold red spots and white patches that make it look like an entirely different, toxic butterfly. Males of the species never wear the second costume — only some daughters do, and it's not a slow gradient between the two looks. It's a hard switch, on or off. Two related Japanese swallowtails independently evolved almost the same switch, and the reason not every female flips it is its own small story about cost and risk. (We'll teach you the Japanese words as we go.)
A costume, not a species
The clearest Japanese example is Papilio polytes, called shiro-obi ageha (シロオビアゲハ, white-banded swallowtail), found through the Ryukyu Islands. Its males and most females share the same pattern: black wings crossed by a white band. But a second female form exists in the same populations, wingtip for wingtip the same size and shape, patterned instead in red, white and black — a close copy of a genuinely poisonous butterfly living in the same habitat. A related mainland species, Papilio memnon, known as nagasaki ageha (ナガサキアゲハ), runs the identical trick against a different toxic model further north. In both cases, this is gitai — 擬態, mimicry — and specifically the kind named after the 19th-century naturalist Henry Bates: a harmless species copying a dangerous one closely enough that predators, having learned to avoid the real thing, avoid the copy too.
What the copied butterfly is actually poisonous with
食草 — shokusō (larval host plant)
Neither swallowtail makes its own poison. The toxin comes from what the caterpillar eats — its shokusō. In the Ryukyus, the toxic model for shiro-obi ageha is Pachliopta aristolochiae, beni-mon ageha (ベニモンアゲハ, crimson rose), whose caterpillars feed on Aristolochia vines (ウマノスズクサ and relatives) and sequester the plant's aristolochic acid alkaloids into their own tissue, keeping the toxin through metamorphosis into the adult. On the mainland, nagasaki ageha's model is Atrophaneura alcinous, jakō ageha (ジャコウアゲハ), which sequesters the same class of toxin from the same plant family. Both toxic models advertise the poison with the same red-white-black warning pattern — and that shared pattern, more than any shared ancestry, is exactly what the mimic females borrow.
How an entire wing pattern flips at once
超遺伝子 — chōidenshi (supergene)
A convincing mimic needs several traits to change together — wing shape, colour placement, pattern edges — and if those traits were controlled by separate genes scattered across the genome, ordinary reshuffling during reproduction would constantly break the combination apart, producing useless half-mimics. Research on shiro-obi ageha (Papilio polytes) traced the entire switch to one region: about 130,000 DNA letters on one chromosome, built around the gene doublesex (dsx) — a gene that, everywhere else in the insect world, mainly decides male versus female body form. In this species a chromosomal inversion locks that whole stretch together so it is inherited as a single unit, a chōidenshi, or supergene: one dominant version (H) switches on the full mimetic pattern, one recessive version (h) leaves the ordinary tiger-stripe pattern in place, and mixing of the two rarely happens because the inversion prevents it. A 2019 study in Communications Biology traced how dsx-H, active only in developing female wings, both switches on the genes that build the red-and-white pattern and switches off the genes that would have built the ordinary one.
A second Japanese species reached the same switch a different way
平行進化 — heikō shinka (parallel evolution)
Nagasaki ageha (Papilio memnon) has its own female-limited mimic form, and a 2018 study led by University of Tokyo researchers found its supergene sitting in a roughly 150,000-letter region on a different chromosome, also built around the dsx gene, also controlling three linked genes as one unit. But the chromosomal inversion that locks the shiro-obi ageha version together is missing here — nagasaki ageha's supergene stays intact through some other, not yet fully identified structural mechanism. Because Papilio polytes and Papilio memnon are close relatives, the two species most likely did not inherit this trick from a shared mimetic ancestor; each one appears to have built its own version of the same dsx-based solution independently. That is heikō shinka, parallel evolution — different routes converging on the same functional answer, because dsx was already sitting in the genome as a switch capable of controlling a whole suite of traits at once, and evolution kept reusing it for that job.
The disguise has a cost, and its value depends on the neighbours
頻度依存 — hindo izon (frequency-dependent)
If mimicry helps butterflies survive, an obvious question follows: why hasn't natural selection made every female a mimic? Two answers have turned up in the research. First, the supergene itself is not free — a 2020 study in Scientific Reports found that shiro-obi ageha females carrying the mimetic dsx-H allele lay fewer eggs, have lower hatching success, and live shorter adult lives than non-mimetic hh females, a real biological cost riding along with the disguise. Second, and more strikingly, a University of the Ryukyus field survey across five islands — Kikaijima, Okinawa Island, Miyakojima, Ishigakijima and Taketomijima — found that the local mimicry rate tracks the local abundance of the toxic model butterfly: where beni-mon ageha is common, mimicry pays off and more females wear the disguise; where the toxic model is scarce, predators have less experience with the warning pattern, mimicry protects less, and non-mimetic females do just as well or better. This is hindo izon selection — an advantage that shrinks as it becomes more common relative to its model — and researchers documented mimicry rates rising over roughly four decades as the toxic model itself spread further through the islands.
So what keeps this whole system from collapsing?
Frequency-dependent selection is what stabilises it: mimicry only works while the toxic model stays common enough that predators have actually learned to leave the warning pattern alone, so the mimic form can never fully take over its own population without undermining the very advantage it depends on. That balance is also why this system is worth watching as Japan's climate shifts. Nagasaki ageha itself has been pushing north for decades — a species once limited to Kyushu and southern Honshu was recorded overwintering in Fukui and Kanagawa by the early 2000s and as far as Fukushima by 2009, tracked by entomologists as an indicator species for regional warming. If a mimic species expands into new territory faster than its toxic model does, or the two separate at different rates, the whole supergene switch stops paying for itself in the newly reached range — a live example of how a single gene can be a powerful adaptation in one place and simply irrelevant in another, depending entirely on who else happens to be living nearby.
Common questions
Q. Why do only female swallowtails in these Japanese species mimic a toxic butterfly, and not males?
A. The mimicry-controlling supergene sits on and around the gene doublesex (dsx), which is already the gene insects use to build male versus female traits differently. The mimetic wing-pattern genes it switches on are only active in female wing development, so the disguise never appears in males regardless of which version of the supergene they carry.
Q. What actually makes the model butterflies (beni-mon ageha and jakō ageha) poisonous?
A. Neither species produces its own toxin. Their caterpillars feed on Aristolochia host plants (shokusō) and sequester the plants' aristolochic acid alkaloids into their bodies, retaining the toxin through metamorphosis into the toxic, warning-coloured adult.
Q. What is a supergene, and why does butterfly mimicry need one?
A. A supergene is a cluster of genes inherited together as a single unit, usually because a chromosomal rearrangement (such as an inversion) prevents them from being separated during reproduction. Mimicry needs several traits — colour, pattern placement, wing-shape details — to change together; a supergene keeps that whole combination intact across generations instead of being broken apart by ordinary genetic reshuffling.
Q. Did Papilio polytes and Papilio memnon inherit the same mimicry supergene from a common ancestor?
A. Research indicates probably not. Both supergenes are built around the same dsx gene, but Papilio polytes relies on a chromosomal inversion that is absent in Papilio memnon, whose supergene appears to be stabilised by a different mechanism. This points to the two closely related species having evolved similar dsx-based mimicry switches independently — parallel evolution rather than shared inheritance.
Q. If mimicry helps a butterfly survive, why do non-mimetic females still exist?
A. Two documented reasons. The mimetic supergene carries a real biological cost — studies found lower egg counts, lower hatching success and shorter adult lifespan in mimetic females. And mimicry's benefit is frequency-dependent: a field survey across five Ryukyu islands found mimicry rates tracked local abundance of the toxic model, so where the toxic model is scarce, predators haven't learned to avoid its pattern and the disguise protects far less, keeping non-mimetic females competitive.
Where to find Japan's wildlife in the wild: our region-by-region wildlife guide. Or explore all 47 prefectures by recorded species in Ikimono Quest, built from open biodiversity data.
Written by naturalists. The science here reflects established research on female-limited Batesian mimicry and its dsx-based supergene control in Papilio polytes and Papilio memnon, aristolochic-acid sequestration from Aristolochia host plants in Pachliopta aristolochiae and Atrophaneura alcinous, and documented frequency-dependent maintenance of the mimetic polymorphism across the Ryukyu Islands. Species identifications are of real, extant Japanese butterflies. Nature is full of exceptions — that's what makes it worth studying.